GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
Accio-Lab
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
Accio-Lab
Model Tree
Accio-Lab/occamy-1.0
Input Modalities
Output Modalities
Supported Functionality
128 records from the train split of CNN/DailyMail, configuration 3.0.0, truncated to 256 tokens. Sixteen separate test records were held out. The exact dataset revision and selection are in data-provenance.json; the quantization configuration is in quantization-recipe.json. No training or fine-tuning was performed.
For expert weights whose observers were not activated, ModelOpt used weight-derived or fused-tensor scaling during export. This limited calibration set does not establish coverage of every expert on every workload. The runtime checks below test the exported checkpoint.
| Check | BF16 | NVFP4 |
|---|---|---|
| Text requests ending normally | 8/8 | 8/8 |
| Strict JSON fixtures | 3/4 | 3/4 |
| Tool-call JSON fixtures | 2/2 | 2/2 |
| Structured tool calls through the API | 2/2 | 2/2 |
| Local add-tool execution and result follow-up | True | True |
| Generated code with independent tests | 2/2 | 2/2 |
| Held-out token NLL | 2.252837 | 2.262056 |
| Synthetic red-image recognition | pass | pass |
Full scalar results are in VALIDATION.json. These are small regression checks, not general benchmark results. The add tool was executed locally and its result returned to the model; the weather case validates the structured call only. Code fences were removed for code execution; JSON checks require raw JSON. No broad vision, audio, video, long-context, or throughput claim is made. The validation context length was 2048.
SGLang 0.5.13.post1, source commit 85fd90072d1a9f2432842b03588f63b745e524e4, FlashInfer 0.6.12, PyTorch 2.11.0+cu130. Native NVFP4 execution requires compatible hardware and kernels.
bash
python -m sglang.launch_server \--model-path Accio-Lab/occamy-1.0-NVFP4 \--quantization modelopt_fp4 --dtype bfloat16 \--moe-runner-backend flashinfer_cutlass \--attention-backend triton --mamba-backend triton \--mm-attention-backend sdpa \--tool-call-parser qwen3_coder --reasoning-parser qwen3 \--context-length 2048 --max-total-tokens 4096 \--max-running-requests 1 --max-mamba-cache-size 4 \--disable-cuda-graph --disable-radix-cache --disable-overlap-schedule
The tests used greedy decoding and chat_template_kwargs={"enable_thinking": false}.
Original model: Accio-Lab, based on Qwen3.6-35B-A3B. Original Apache-2.0 license is included. This repository contains quantized Occamy weights, not the separately published NVIDIA quantization of the original Qwen model.